Maps in the Head and Maps in the Hand Kenny Skagerlund ([email protected]) Department of Behavioral Sciences, Linköping University SE-581 83 Linköping, Sweden David Kirsh ([email protected]) Cognitive Science Department, University of California, San Diego La Jolla, CA 92093, USA Nils Dahlbäck ([email protected]) Department of Computer and Information Science, Linköping University SE-581 83 Linköping, Sweden Abstract Lawton (1994; 1996) found that people tend to report using one either an orientation strategy or a route strategy Using the perspective of situated cognition we studied how people interact with a physical map to help them navigate when navigating, but not both. Orientation strategies are through an unfamiliar environment. The study used a mixture cognitive processes that use survey knowledge, the umbrella of cognitive ethnography and traditional experimental term for world-centric relations. When a subject thinks in an methods. We found that the difference between high and low allocentric reference frame using global attributes of a performing navigators showed up in the speed they completed terrain such as cardinal directions, and Euclidean their task and also in the way they use maps. High performers positioning of landmarks, they are using orientation or plan routes using a survey method whereas low performers use a route strategy. We suggest that when people are given a survey strategies (Prestopnik & Roskos-Ewoldsen, 2000) task that does not match their cognitive style they try to for wayfinding. Route strategies, by contrast, are based on transform the task to better suit their cognitive abilities and an egocentric frame of reference, where paths are defined as cognitive style. those throughways available from where the subject is at the Keywords: Map use, navigation, wayfinding, situated moment. cognition, spatial cognition. Research on spatial knowledge acquisition and navigation has mostly been confined to strict laboratory settings and Introduction virtual environments (e.g. Allen, Kirasic, Dobson, Long, & Beck, 1996; Montello, 1993), or to environments where the Interest in human spatial cognition and navigational subject is led along a fixed route in an urban area (Kato & capacity has a long history, ranging from the pioneering Takeuchi, 2003). In these studies maps have been used work of Siegel & White (1975) to contemporary primarily as a diagnostic tool to reveal the subject’s internal contributions by Montello (1998; 2005) and Hegarty et al. representation. For instance, a subject might be asked to (2002; 2006). Spatial cognition is concerned with how sketch the route she followed, marking down all the people represent space and navigate through it. (Montello, landmarks she can recall. (Liben, 2010). Little or no 2005). In the “classical” view, knowledge, from the attention has been paid to the actual practices of subjects perceived environment, is represented as a cognitive map when they use maps to navigate. (Tolman, 1948, Galotti, 2008). Siegel & White (1975) A map, if properly used, is an artifact that extends a distinguished three types of knowledge involved in forming person’s survey knowledge (Montello, Hegarty and and using cognitive maps: i) landmark knowledge, ii) route Richardson 2004). It behaves in the same way as an internal knowledge, and iii) survey knowledge. Landmark knowledge map except that it is external. Because we interact with is information about the particular features at a location. internal and external representations differently, however, it Route knowledge is information about specific pathways for is worth examining in detail the diverse ways that people moving from one location to another; it may be coded interact with maps. Do all subjects rotate maps? When, procedurally or declaratively. Survey knowledge is metric why? How do they gesture? Do they point on the map and information about the relative location and estimated then to the world? How often do they glance at a map? distances between landmarks, the very thing captured in a When? standard map, showing the location of all paths and features To study the practices of map use we videoed subjects in a Euclidean plane. All this work investigates the using a map of UCSD campus as they found their way from representational architecture of internal spatial a starting location to a goal location. In the analysis we representations, focusing on questions such as whether divided our subjects into two groups – route-based cognitive maps are map-like in nature or more like nodes in navigators, and survey-based navigators – using the well- a graph representation. known measures developed by Lawton (1994). We report here on our findings and offer an explanation of the results 2339 in line with the ideas of situated, distributed and embodied The Santa Barbara Sense of Direction Scale (SBSOD) cognition. was used to measure the participants’ sense of orientation, Clark (2008) has coined a term, or principle, to label this or the awareness of location or orientation. This instrument type of interaction in which epistemic actions (Kirsh & is a self-report measure which has been found to predict Maglio, 1994) can be incorporated – The Principle of objective measures of these abilities, such as dead reckoning Ecological Assembly (PEA). This principle states that “the (Hegarty et al., 2002). This instrument has proved to be canny cognizer tends to recruit, on the spot, whatever mix of internally consistent and has sufficient test-retest reliability. problem-solving resources will yield an acceptable result The SBSOD is highly correlated with measures of spatial with a minimum of effort” (Clark, 2008, p.13). But, given knowledge acquired from direct experience in the the existence of individual differences in cognitive styles, it environment, and Hegarty et al. (2002) has shown that it is is not obvious that “a minimum of effort” means the same related to knowledge that involve orienting oneself within thing for all people. A specific instance of this is the the environment. difference between persons who have a preference for a The Wayfinding Strategy Scale (Lawton 1994) is a survey route strategy or a survey strategy. Another aim of the that measures to what extent a person depends on strategies present work is to study how map use might differ between relying on survey knowledge or route knowledge people depending on their preferred navigation strategies. respectively. The survey contains 14 items of the sort of Since we are bridging or trying to relate two different propositions that participants have to grade the degree of research traditions, we address our research questions using agreement along a 5-point Likert-type scale. a combination of the experimental methods traditionally To measure dead reckoning a pointing task was used. At a used in research on navigation and wayfinding, with number of places the participants were asked to point in the cognitive ethnography used in research on situated and direction of an unseen landmark; a traditional compass was distributed cognition. used to assess the participants’ error in this task. Method Procedure The study was undertaken on the University of California, The study itself was divided into two separate sessions. The San Diego campus. UCSD is sufficiently complex and first session was a pretest, where participants filled out covers a large enough area to be challenging for most electronic counterparts to the physical instances of SBSOD navigators unfamiliar with the campus. It can also be and the wayfinding strategy scale over the internet. The considered representative of an urbanized area surveys could be completed at any time the participants 17 participants were recruited using craigslist, which is an wished from the moment of agreement of participation in online ad-service where ads can be placed for a fee. The the study to the day when the experiment session began. The participants were between 20 and 58 years, Mean =32.1 (SD questionnaire was filled out prior to the experiment session, = 13.23), 8 female, 9 male; they were unfamiliar with the which was of vital importance as to ensure validity. If it UCSD campus. To eliminate vision as factor in would have been completed after the experiment trial, there performance they had to have 20/20 vision – with or without would have been a possible risk that participants took into corrective lenses or glasses. account their recent navigational performance, and thus Participants were asked to find their way from a starting affecting the self-assessment. point to goal location. Three different start-goal pairs were On arriving for the second session, a consent form was used. These pairs were chosen and evaluated during a pilot filled out by each of the participants, a parking permit was study, where they were determined to be equally hard. paid for and given to them if needed, and they were then Criteria for hardness were the number of salient landmarks, told to step into a car for transportation to another location. the density of buildings throughout the area, length (air From this moment on, the experimental session had distance), visual access. By using start-goal pairs that officially started and they were instructed to try to pay overlapped and crossed through the campus center, the attention to where they are located in the world from that environmental features and vistas were as equivalent as point onwards. The participant was dropped off at one of the possible, leading us to infer they were equally complex. marked drop-off locations where they were picked up by another experimenter. On site of the drop-off point the Materials equipment was set-up, which included mounting the The materials used in this study included the official visitor headcam on the participant and getting a stable GPS signal. map of the UCSD campus, which was handed to the The participant was then told to estimate and point into the participants and used throughout the navigation task.